Sure--we can use our fantastic powers of generalization and comprehension of complex situations on our moral instincts the same way that we can on our other instincts, in order to find a mutually-mostly-agreeable universalization of the instincts. Nothing wrong with bootstrapping!
(Our general-purpose ability to do this also evolved. And our genes encode for general-purpose ability, including the formation of minds that can adopt culture, precisely because having behavioral flexibility on shorter timescales than is accessible by evolution is profoundly adaptive.)
But if you accept that, then I don't understand your previous objection. In particular, the universalization of pro-social behavior rules out rape under any remotely familiar situation, so I don't know why you keep bringing it up.
We evolved a sense of ought and it's the only "ought" there is. Evolution gave us "ought", manufactured, in essence, out of the demands of survival as a social creature. You don't get an ought from an is through logic, you get it through physical construction. Once you build one, you have an oracle that you can consult for answers to ought-questions; if you query cleverly, you can get at what the effectively foundational oughts are. But you can't just say, "well, naturalistic fallacy--just because it's there doesn't mean we have to use it". In this case, we actually do have to use it because there is no other source for motivational content as opposed to analytical content.
"Ought" adds both an implicit preference-scoring function to states of affairs, and a coupling between passive analysis of preference and behavioral control. If you don't have these things, "ought" doesn't work. We do have this, so it works, but we got it because it's there, not because one can reason to it from nothing. It's essentially a sort of qualia--you don't reason your way to what red feels like; redness just is. Nobody argues about the "naturalistic fallacy of qualia".
But the qualia of ought--what it feels like to ought to do something or that something ought to happen--is...just...qualia. Without the qualia, there would be no imperative, just listings of action-and-outcome-score and...whatever! So, we have a bunch of mappings to scores. Whatever!
(I am speaking as if it's a distinct, separable module. Brains generally don't work that way; things tend to be integrated in a not-wholly-separable fashion. But for purposes of the argument, it's easier to consider a distinct separable module first, and then fix up any logical errors introduced by postulating the separation as stronger than it really is.)
Once we have ought-instincts, we can use our standard tricks of generalization and explicit formulation of goals to sculpt our instincts to be more fit-for-purpose (circularly, because the ought-instinct plays a central role in defining the purpose, but again, there's nothing else there, so...). Bootstrapping is totally fine. You just need boots to start with.